Our DNA is written in Objective-C

Dr. Touch's incredible POI Weight Loss Formula

I’ve been busily hacking away on this for the past week and now I have reached the level where I can formally announce this new addition to Dr. Touch’s Parts Store.

I was approached a week ago by somebody who suggest this to be sold as an inexpensive component and I immediately was intrigued by the problem. Imagine that you have an iPhone app which displays lots of PINs in a very restricted area. Naturally shops would be very close to each other causing the resulting annotation PINs to overlap. DTClusterMaker, the new component, aims to address this problem.

Without it your screen might look like this:

After a quick optimization with DTClusterMaker this mess turns into something much ligther, much nicer to look at:

I hope that you agree that the second view would be the preferred variant. This way you can still see the name of the city in question: München (= Munich).

DTClusterMaker employs a very efficient method to spatially sort and cluster closeby POIs. In the presented demonstration app you can explode any of the purple pins into it’s original POIs. You get two different algorithms to choose from: The perfect one, that tends to slow down if you feed it hundrets of POIs. And it’s ultra-fast cousin which works by “boxing” the POIs. It’s so fast in fact, that I am thinking of letting it run multiple times with a shifted grid and then choose the output that gives me the smallest number of POIs. If you have hundreds of POIs then you won’t see a difference.

DTClusterMaker is available now for you to purchase at an introductory price of 100 Euros. It still needs a bit of work in getting the interface to your code ironed out, but as with all components your access to the SVN repository where it is being developed insures that you get all those improvements for free. Also I am depending on your requirements to figure out the most elegant way to interact with the ultimate POI weight loss formula. 🙂

Categories: Parts

Leave a Comment

%d bloggers like this: